Papers

Low blood pressure and depression in older men: a population based study

BMJ 1994; 308 doi: https://doi.org/10.1136/bmj.308.6926.446 (Published 12 February 1994) Cite this as: BMJ 1994;308:446
  1. E Barrett-Connor,
  2. L A Palinkas
  1. Department of Family and Preventive Medicine, University of California, San Diego, CA 92093-0607, USA
  1. Correspondence to Dr Elizabeth Barrett -Conor.
  • Accepted 1 November 1993

Abstract

Objective : To determine if an association exists between low blood pressure and depressive symptoms in older men living in the community.

Design : Cross sectional, population based study.

Setting : Town of Rancho Bernardo, California, United States.

Subjects : 846 men aged 60-89 years. Comparisons between hypotensive, normotensive, and hypertensive groups were limited to 594 men not taking drugs for hypertension.

Main Outcome Measures : Mean scores on Beck depression inventory and prevalence of scores >= 13. Results - Men with diastolic blood pressure <75 mm Hg had significantly higher depression scores (mean scores 6.35 v 4.96; P<0.001) and more categorical depression (7.6% v 1.8% with scores >=13; P<0.01) than men with diastolic blood pressure levels between 75 and 85 mm Hg. Men with diastolic blood pressure levels >85 mm Hg had higher depression scores than men with intermediate blood pressure levels (mean scores 5.85 v 4.96; P<0.05). Men with diastolic hypotension scored significantly higher on both affective and somatic item subscales of the Beck depression inventory and on individual measures of fatigue, pessimism, sadness, loss of appetite, weight loss, and preoccupation with health. Low diastolic blood pressure was a significant predictor of both mean depression score and prevalence of categorical depression, independent of age and change in weight since the baseline visit. The presence of several chronic diseases was associated with depressed mood and higher blood pressure but not with low blood pressure.

Conclusion : The association of relatively low diastolic blood pressure with higher depressive symptom scores and rates of categorical depression was independent of age or weight loss. Since fatigue is a prominent symptom of depression, any association of low blood pressure with fatigue could reflect depressive disorders or clinically important depression.

Clinical implications

  • Clinical implications

  • The dangers of high blood pressure are well recognised, but the clinical significance of low blood pressure is controversial

  • Patients with low blood pressure often complain of being tired and crying easily

  • In this study, older men with low diastolic blood pressure (<75 mm Hg) were more likely to be depressed and experience fatigue than men with normal (75-85 mm Hg) diastolic blood pressure

  • This association was not explained by age, weight loss, presence of chronic disease, or use of antihypertensive or antidepressant drugs

  • As fatigue is a prominent symptom of depression, clinically important depression should be considered in patients with low blood pressure who complain of fatigue.

Introduction

The dangers of high blood pressure are well recognised, but the clinical importance of low blood pressure is controversial. Several recent large studies reported an increased prevalence of fatigue, crying, or psychological dysfunction associated with low blood pressure.*RF 1-3* Although a hypotensive syndrome manifested by subjective symptoms has been accepted in Europe, it has been dismissed in the United Kingdom and United States.4,5

Depressive symptoms have been described in hypertensive subjects and have been variously ascribed to antihypertensive drugs or the effect of labelling of patients.6,7 To our knowledge no one has previously used a standard depression inventory to study the relation of depressive symptoms to low blood pressure in a population. We report here a cross sectional study of low blood pressure, depression, fatigue, and other somatic complaints in 594 older men living in the community.

Methods

In 1972-4, 82% (5052) of 6155 eligible adult residents of Rancho Bernardo, California, a geographically defined community, participated in a baseline evaluation which included a measurement of weight with subjects wearing light clothing and no shoes. The population is stable, white, and relatively homogeneous for social class and education. In 1984-7, all 1070 surviving non-institutionalised men aged 60 and older were invited to a clinic visit; 79% participated, and these 846 men form the basis of this report.

Participants completed a standardised medical questionnaire, which included questions about cigarette smoking, alcohol use, and exercise. Each subject was asked whether he had ever been diagnosed as having or had been treated for any of 17 common conditions (heart disease; high blood pressure; lung, liver, kidney, gallbladder, or prostate disease; cancer; arthritis; gall stones; thyroid disorders; stomach or duodenal ulcers; diverticulitis; emphysema; chronic constipation; stroke; or hip fracture). Current use of drugs was recorded by a trained interviewer and validated by drugs and prescriptions brought to the clinic for that purpose. Weight was measured according to the same protocol used at baseline. Blood pressure was measured twice by technicians certified for the hypertension and detection follow up programme protocol,8 using a standard sphygmomanometer in seated resting subjects. The means of the two systolic and diastolic readings were used for analysis.

Information on depressed mood at the time of the visit was obtained from responses to 18 of the 21 items of a self administered Beck depression inventory.9 In accordance with the criteria described by Shrout and Yager,10 three of the original 21 items (guilt, expectation of punishment, self hate) were excluded from the questionnaire in an effort to reduce the length of the scale without compromising its reliability in this population. Total scores were proportionally adjusted to correspond to scores and cut points established for the full 21 item scale. In addition, two subscales comprised of affective or non-physical and somatic or physical symptoms were calculated on the basis of the classification of items described by Plumb and Holland.11

The original 21 item Beck depression inventory and a shortened 13 item version have been tested and validated in elderly subjects. Internal consistency was shown in the Rancho Bernardo cohort by a Cronbach's (alpha) of 0.69. A cut point of 13 was used to define cases of mild to severe depression among study subjects on the basis of its use as a reliable indicator of mild to severe depression in populations at risk for chronic disease and disability.12

Statistical methods

Men with normal blood pressure (defined before any analyses as 75-84 mm Hg diastolic or 120-139 mm Hg systolic) were used as the comparison group in all bivariate analyses. Risk factor distribution and mean depressive symptom scores and prevalence rates were determined separately for men not taking drugs for blood pressure. Mean depressive symptom scores were adjusted for age and compared by using an analysis of covariance procedure.13 Prevalence rates were adjusted for age with the direct method,14 with the total sample comprising the standard population. The Mantel-Haenszel age adjusted X2 test was used to compare the age adjusted prevalence rates.15 However no adjustment was made for multiple comparisons, and caution should be exercised in interpreting results.

Non-parametric tests (Spearman's r) were used to assess the association between depressive symptoms, chronic diseases, and other factors. Each independent variable found to be significantly correlated with depressive symptoms and diastolic blood pressure was then entered into a stepwise multiple regression model.

Results

Among the 846 men aged 60-89 years in this cohort, 252 (30%) were using antihypertensive drugs at the time of the study. Blood pressure was not associated with depression score or categorical depression in these men, and they were excluded from further consideration as the purpose of this study was to examine the relation of naturally occurring low blood pressure to depressed mood and fatigue.

Over a third (255/594) of the men not receiving antihypertensive treatment had a diastolic blood pressure <75 mm Hg. The prevalence of both diastolic hypotension and depressed mood increased with age (r=0.18, P=<0.001 and r=0.15, P<0.001, respectively). As shown in Table I, in age adjusted analyses men with low diastolic blood pressure had significantly higher depression scores (P<0.001) and more categorical depression (P<0.01) than men who had intermediate diastolic blood pressure (75-85 mm Hg). Men with diastolic blood pressures >85 mm Hg also had higher depression scores (P<0.05) than men with intermediate levels. Low systolic blood pressure, defined as <120 mm Hg, was much less common and was only marginally associated with depression symptom scores or depression (P<0.1).

TABLE I

Relation of mean scores on Beck depression inventory scores (adjusted for age) and prevalence of scores >=13 to diastolic and systolic blood pressure in men aged 60-89 not receiving drug treatment for high blood pressure Rancho Bernardo, California

View this table:

As shown in Table II, men with diastolic hypotension scored significantly higher on both the affective and somatic item subscales of the Beck depression inventory and on individual measures of fatigue, pessimism sadness, loss of appetite, weight loss, and preoccupation with health. Men with systolic hypotension did not differ in overall depressive symptom scores although they did score significantly higher in fatigue, pessimism, and lack of satisfaction.

TABLE II

Relation of mean Beck depression inventory scores (adjusted for age) by question and blood pressure, 60-89 year old men, Rancho Bernardo, California, 1984-7

View this table:

As shown in Table III, only age and weight loss since the baseline visit were associated with both depressive symptom scores and low diastolic blood pressure. Several other plausible covariates of depression and low blood pressure did not explain these associations. Thus, although depression scores were significantly higher in men who reported chronic conditions (emphysema, arthritis, gall stones, ulcers, diverticulitis, constipation, cancer, stroke, and prostate surgery), constipation was the only chronic condition associated with low blood pressure. As Table III shows, the number of chronic conditions was associated with the mean score on the Beck depression inventory but not with low blood pressure.

TABLE III

Correlations between Beck depression inventory score, diastolic blood pressure, and demographic and health status measures, 60-89 year old men, Rancho Bernardo, California, 1984-7

View this table:

The use of antidepressant drugs also did not explain the association. Only 35 of the 594 men were taking antidepressants. The prevalence of categorical depression was associated with antidepressant use (17.1% (6/35) v 3.8% (20/528) not taking antidepressants; X2=13.3, P<0.001) but was not associated with the prevalence of low systolic or diastolic blood pressure. The Beck depression inventory score was significantly associated with use of antidepressant drugs, but low blood pressure was not associated with use of antidepressants (table III). Similarly, lack of regular exercise was associated with depressive symptom score but not with low blood pressure.

Low diastolic blood pressure, age, and change in weight since the baseline visit were entered into a stepwise multiple regression model to test their independent effects on depressive symptom scores. Each of these three variables was significantly and independently associated with the depressive symptom score (table IV). Additional evidence for an independent association between low diastolic blood pressure and depressive symptoms was provided by logistic regression analysis to test the effects of low diastolic blood pressure, age, and change in weight since the first visit on the likelihood of a Beck depression inventory score >=13. The log adjusted odds ratio of categorical depression in men with low diastolic blood pressure compared with men with a diastolic blood pressure >=75 mm Hg was 2.92 (95% confidence interval 1.25 to 6.86). Age (60-74 v 75-89 years; odds ratio 2.33 (0.94 to 5.73) and change in weight (loss or no change v gain; 1.41; (0.81 to 3.67) were not significant independent predictors of categorical depression in this model.

TABLE IV

Regression of Beck depression inventory score on low diastolic blood pressure,age, and weight loss since baseline visit, 60-89 year old men, Rancho Bernardo, California, 1984-7

View this table:

Discussion

In this cohort, men who had relatively low diastolic blood pressure in the absence of drugs for hypertension had significantly higher symptom scores for depression and more categorical depression than men with intermediate levels of blood pressure. The prevalence of both depression and hypotension increased with age, but age did not explain the association. This association was also not explained by weight loss, other chronic disease, or the use of antidepressants. Beck depression scores and categorical depression were more strongly associated with low diastolic than low systolic blood pressure, but low systolic pressure was present in only 22 men who did not also have low diastolic blood pressure.

The individual items on the Beck depression inventory were disaggregated and fatigue and other somatic items were looked for as determinants of depression scores; fatigue, loss of appetite, weight loss, and preoccupation with health (somatic complaints) were found to be significantly higher in men with diastolic hypotension. Fatigue was also significantly higher in men with systolic hypotension. Overall, eight (of eight) somatic and eight (of 10) affective items on the Beck depression inventory were higher in hypotensive men, and seven of these differences were significant at the 5% level.

Although low blood pressure has long been thought to be a cause of tiredness and dizziness, relatively few studies with appropriate comparison groups have been conducted outside of clinical settings. Their findings are generally concordant with those of the present study. Bengtsson et al reported an increased prevalence of fatigue, dizziness, and readiness to cry in a population based study of 1302 women with a systolic blood pressure below 120 mm Hg.1 Wessley et al found a linear relation between tiredness and low blood pressure unexplained by psychological illness in 7382 adults aged 18 and older.2 In a study of 10314 civil servants aged 33-55 Pilgrim et al also reported an excess of tiredness in those in the lowest fourth of systolic blood pressure; they concluded that this difference was entirely explained by psychological dysfunction but found no evidence of a particular specific condition.3 Interestingly, Pemberton, who most effectively challenged the concept that low blood pressure was associated with morbidity, found a significant negative trend for both systolic and diastolic blood pressure in women with tiredness.5 Unlike the present study, these earlier studies did not examine elderly people and did not examine depression by using a standard instrument. Other attempts to relate blood pressure to depressive symptoms have focused on patients with hypertension6,16 or psychiatric illness.17

No mechanism for the association of low blood pressure with depressed mood is known. We cannot exclude the possibility that low blood pressure leads to fatigue, anorexia, and depressed mood. Since fatigue is a prominent symptom of depression, clinically important depression should be considered in patients with low blood pressure and fatigue. This association is found with sufficient consistency to be credible and to merit further attention.

This study was supported by National Institute of Aging grant No R37 AG07181-05 and National Institute of Diabetes, Digestive, and Kidney Diseases grant No R01-DK31801.

References

View Abstract